786  Prof.  R.  W.  Wood  on 
For  these  cases  there  would  be  a  marked  increase  in  the 
intrinsic  intensity  as  the  surface  of  the  source  was  fore- 
shortened, but  it  would  not  be  as  great  as  in  the  case  of  a 
gas-flame,  or  the  fluorescent  prism  which  we  have  considered. 
Lommel  observed  this  increase  in  the  apparent  intensity 
by  immersing  a  vessel  containing  the  fluorescent  medium  in 
a  tank  of  water,  and  observing  the  luminous  surface  at  dif- 
ferent angles.  Lommel  considered  his  fluorescent  medium 
as  partially  absorbing  for  the  fluorescent  radiations,  in  which 
respect  it  differs  from  the  crown  glass  used  in  the  present 
experiments,  which  is  almost  perfectly  transparent  to  the 
blue  fluorescent  light  which  it  emits. 
From  what  has  been  already  stated  it  is  clear  that  the 
intensity  of  the  fluorescent  surface,  when  viewed  from  the 
air  side  at  varying  angles,  is  profoundly  influenced  by  re- 
fraction. We  should  naturally  expect  radiations  originating 
at  a  small  depth  within  the  medium  to  be  refracted  on 
emergence  into  the  air,  but  it  is  not  quite  clear  what  would 
happen  to  radiations  leaving  the  surface-layer  of  molecules. 
Millikan  came  to  the  conclusion,  from  a  study  of  the 
polarization  of  the  fluorescent  light,  that  these  radiations  are 
refracted  in  the  same  way  as  radiations  coming  from  a  finite 
depth  within  the  medium.  In  the  case  of  fluorescence  pro- 
voked by  light,  it  is  probable  that  the  radiating  layer  is  of  con- 
siderable thickness  as  compared  with  the  molecular  diameter, 
for  the  light  undoubtedly  penetrates  to  the  depth  of  a  wave- 
length or  two.  It  seemed  possible  that  an  effect  more  super- 
ficial in  its  nature  might  be  obtained  by  exciting  the 
fluorescence  with  cathode  rays.  A  glass  sphere  was  mounted 
in  a  vacuum-tube  in  front  of  the  cathode,  but  it  wTas  found 
that  the  appearance  was  the  same  as  with  ultra-violet  illumi- 
nation. The  tube  was  so  constructed  that  the  sphere  could 
be  bombarded  either  with  cathode  rays  or  canal  rays ;  and  it 
was  found  that  if  the  latter  were  caused  to  play  upon  the 
surface,  a  phosphorescence  was  obtained  which  did  not  obey 
Lambert's  law,  for  the  light  appeared  much  more  intense 
around  the  edge  of  the  sphere,  where  the  surface  was  con- 
siderably foreshortened.  This  light,  as  is  well  known,  shows 
the  D  lines  strongly,  and  comes  without  doubt  from  a  shallow 
layer  of  gas,  which  clings  close  to  the  surface.  It  is  possible, 
however,  that  in  this  case  we  have  an  emission  of  light  from 
the  surface  molecules  alone  ;  in  other  words,  the  effect  which 
was  sought  for  in  the  cathode-ray  phosphorescence.  The 
canal  rays  have  without  doubt  much  less  penetrating  power, 
and  may  excite  a  purely  superficial  phosphorescence.  My 
feeling    about    the   matter,   however,  is   that    the    luminous 
